Linear Solvation Energy Relationships. 22. Some ... - ACS Publications

Apr 28, 2015 - O for aromatic solvents. Where type A hydrogen bonding (solvent donor/solute acceptor) also influences the XYZ term, the relationships...
0 downloads 0 Views 537KB Size
250

Inorg. Chem. 1983, 22, 250-254 Contribution from the Department of Chemistry, University of California, Irvine, California 927 17, and Naval Surface Weapons Center, White Oak Laboratory, Silver Spring, Maryland 20910

Linear Solvation Energy Relationships. 22. Some Further Examples of the Importance of the d6 Term R. W. TAFT*l8 and MORTIMER J. KAMLET*lb Received May 4, 1982

Examples are shown of attempts at correlation of solvent effects by the LSER equations (a) XYZ = XYZo + ST* and (b) XYZ = XYZ, + S(T* + dS). Where the absolute magnitude of the d term is relatively small, correlations by equation a may appear to succeed, even though equation b is the more correct equation. For larger d values, however, correlations are apt to fail unless equation b is used. Examples described are the "Se NMR shifts of dimethyl selenide, the 125TeNMR shifts of dimethyl telluride, and the lowest energy band in the electronic spectrum of tungsten tetracarbonyl 1,lO-phenanthroline. We have shown earlier2 that, when hydrogen-bonding interactions are excluded by a judicious choice of solvents and reactants or indicators, correlations of many types of reactivity parameters and physicochemical properties can take either of two forms: (a) For all electronic spectral transitions so far considered which are shifted bathochromically with increasing solvent dipolarity, with all solvents considered together, and for other properties if families of solvents with similar polarizability characteristics are treated separately (e.g., only non-chlorinated aliphatic solvents, only polychlorinated aliphatics, or only aromatic solvents), the form of the linear solvation energy relationship (LSER) is given by eq 1, where X Y Z = XYZO SA* (1)

+

A* is a dipolarity/polarizability parameter that measures the solvent's ability to stabilize a charge or a dipole by virtue of its dielectric effecte3 (b) For the other types of properties, if all solvents need to be considered together, the preferred from of the LSER becomes eq 2, where 6, a "polarizability X Y Z = XYZ, S ( A * + d6) (2)

+

correction term", is 0.0 for non-chlorinated aliphatic solvents, 0.5 for polychlorinated aliphatics, and 1.O for aromatic solvents. Where type A hydrogen bonding (solvent donor/solute acceptor) also influences the XYZ term, the relationships corresponding to eq 1 and 2 are given by eq 3 and 4. Where X Y Z = XYZO sx* + ua (3)

+ d6) + ua XYZ = XYZO + SA* + b@ XYZ = XYZO + S(A* + d6) + b@

XYZ = XYZO

+

+

S(A*

(4)

(5) (6)

type B hydrogen bonding (solute donor/solvent acceptor) plays a part, the corresponding equations are eq 5 and 6, where CY is a measure of solvent HBD (hydrogen bond donor) acidityhs4 and @ a measure of solvent HBA (hydrogen bond acceptor) b a s i ~ i t y . The ~ ~ ~XYZ's that have been correlated by eq 1-6 include positions and/or intensities of maximal absorption in IR, UV/visible, ESR, and N M R spectra, N M R coupling constants, free energies of solution or of transfer between solvents of dipolar solutes, and logarithms of reaction rate and equilibrium constants and fluorescence lifetimes. (a) University of California. (b) White Oak Laboratory. (a) Taft. R. W.: Abboud. J.-L. M.: Kamlet, M. J. J. A m . Chem. SOC. 1981, 103, 1080. (b) Kamlet, M. J.; Abboud, J.-L. M.; Taft, R. W. Prog. Phys. Org. Chem. 1981, 13, 485. For most non-chlorinated aliphatic solvents (specifically excluding dioxane and hexamethylphosphoramide) a* values are very nearly proportional to molecular dipole moments: Abboud, J.-L. M.; Kamlet, M. J.; Taft, R. W. J. Am. Chem. SOC.1977, 99, 8321. Taft, R. W.; Kamlet, M. J. J. Chem. Soc.,Perkin Trans. 2 1979, 1723; J . Am. Chem. SOC.1976, 98, 2886. (a) Kamlet, M. J.; Jones, M. E.; Taft, R. W.; Abboud, J.-L. M. J. Chem. Soc., Perkin Trans. 2 1979, 342. (b) Kamlet, M. J.; Taft, R. W. J. Am. Chem. SOC.1976, 98, 311.

0020-1669/83/1322-0250$01.50/0

Frequently, although eq 2,4, or 6 is the statistically favored form of the LSER, correlations of type (b) properties by eq 1, 3, or 5 have still shown seemingly acceptable statistical measures of goodness of fit (Le., correlation coefficients, r, greater than 0.95). This is usually the case where the d value in eq 2, 4, or 6 is relatively small (d < 0.15) and the spread of the XYZ values relatively large. Thus, we have reported6 that, although eq 2 is, indeed, the preferred form of the LSER, solvent shifts of most infrared spectral bands are adequately correlated by eq 1 (of r values for 25 correlations, all were greater than 0.90, 23 were greater than 0.95, and 21 were greater than 0.97). Where the "polarizability correction term" is larger ( d > 0.30) and the XYZ spread relatively smaller, however, attempts at fitting experimental data to eq 1, 3, or 5 have sometimes resulted in seeming lack of correlation. 77Se NMR Shifts of Dimethyl Selenide As an example of the latter type of experience, Valeev, Kalabin, and Kushnarev' have reported poor correlations of solvent effects on the 77SeN M R shifts of (CH3)2Sewith the solvatochromic parameters a*,CY,and 0. Insufficient CY values were said to be available for a correlation according to eq 3 (the Russian workers were evidently not aware that non-HBD solvents have a values of zero), and the multiple-parameter correlation with A* and @ according to eq 5 had a correlation coefficient of 0.502 ( n = 13). We wish now to demonstrate the salutary effect of including a d6 term in such correlations. 77SeN M R A6 values have been obtained for 5 mol 5% solutions of MelSe in 19 solvents for which the solvatochromic parameters are known.7 The shifts are given relative to neat Me,Se as an external reference, and the authors have shown that the contributions to these shifts from the volume diamagnetic susceptibilities and the magnetic anisotropies of the solvents are no more than 1-2 ppm (Le., differences in these contributions are negligibly small). The results are assembled with solvent A* and a values in Table I and plotted against a* in Figure 1. It is seen in the plot that, if all solvents are considered equally, linear correlation is, indeed, quite poor. However, when the aliphatic and aromatic solvents are considered separately, reasonably good linear progressions become evident. Further, the data points for the aliphatic HBD solvents nitromethane (32), methanol (lo$ and acetic acid (201) appear to be collinear with those for the non-hydrogen-bonding aliphatic solvents, suggesting that the 77SeNMR shifts are little, if at all, influenced by type A hydrogen bonding by these proton donors. The least-squares linear regression equations are (a) for ten non-chlorinated HBD and non-HBD aliphatic solvents eq 7, A6(77SeNMR) = 3.2 - 1 7 . 5 ~ *ppm*

(7)

(6) Kamlet, M. J.; Taft, R. W. J. Chem. SOC.,Perkin Trans 2 1979, 331. (7) Valeev, R. B.; Kalabin, G. A.; Kushnarev, D. F. Zh. Org. Khim. 1980, 16, 2482.

0 1983 American Chemical Society

Inorganic Chemistry, Vol. 22, No. 2, 1983 251

Linear Solvation Energy Relationships Table I. Solvent Effects on Properties Studied

NMR shifts, p p m no!

2 3 6 7 8 14 15 16 17 18 20 21 23 24 26 29 30 31 32 41 53 57 58 61 63 90 97 104 105 109 201 a

solvent

ff*

a

cyclohexane triethylamine carbon tetrachloride diethyl ether toluene benzene chlorobenzene 2-butanone anisole acetone 1,2-dichloroethane methylene chloride dimethylformamide pyridine hexamethylphosphoramide dimethyl sulfoxide chloroform nitrobenzene nitromethane cyclohexanone mesitylene phenylacetonitrile acetophenone 1,2-dimethoxyethane N,N-dimethylaniline cumene fluorobenzene ethanol methanol benzyl alcohol acetic acid

0.0

0.0 0.0 0.0 0.0 0.0 0.0 0.0

Solvent numbering is the same in all papers of this series.

0.14 0.29 0.27 0.54 0.59 0.71 0.67 0.73 0.71 0.81 0.83 0.88 0.87 0.88 1.00 0.58 1.01 0.85 0.76 0.41 0.99 0.90 0.53 0.90 0.41 0.62 0.54 0.6 0 0.98 0.64

"Se

4.5 0.4 2.1 -2.8

10-3umax(I), cm-'

5.9 0.9 -0.2

19.05

0.0 -3.2

-3.1

-9.8

-12.7

-1.2 -11.9 -5.5 -9.4 -14.9 -7.6 -5.4 - 13.4

-14.9 - 18.6 -9.8

19.23 19.57 20.62 20.00 21.05 20.80 20.08 21.46

0.05

0.0 0.09

0.0 0.25

0.0 0.0 0.0 0.0 0.35

0.0 0.23

0.0 0.0

-26.4 -6.3 -7.1 -20.6

0.3

nydb nydb

21.83 19.49 20.75 21.65 20.45

5.8

-7.0 20.70 20.34

0.0 0.0 0.0 0.0

-4.2 1.2 -5.9

0.86 0.98 0.43 1.01

-8.3

20.20 20.79 20.70

-11.2

-6.7

Not yet determined.

with r = 0.975 and sd (the standard deviation) = 1.5 ppm, and (b) for six aromatic solvents eq 8, with r = 0.927 and sd

9

A6(77SeNMR) = 3.6 - 9.8a* ppm = 1.1 ppm. From eq 7 and 8, the d term in eq 2 is calculated to be -0.42,9 and the "all-solvents" correlation for the 18 solvents of Table I (excluding chloroform)lOJ1is given by eq 9, with r = 0.9591° and sd = 2.1 ppm. A plot of A6(exptl)

A6(77SeNMR) = 2.2 - 15.9(?r* - 0.426) ppm

lz5Te

x6

(9)

against ?r* - 0.426 is shown in Figure 2, where it is seen that, in contrast with the poor correlations reported by Valeev and mworkers, inclusion of the d6 term leads to a quite acceptable linear regression equation. The correlation equations indicate, rather convincingly, that solvent effects in the 77SeNMR shifts of dimethyl selenide are influenced primarily (or exclusively) by dipolarity and little, if at all, by HBD acidity (except possibly in the case of chloroform)" or polarizability (vide infra). When it is taken with eq 7 and 9, the A6 value of 0.0 for neat Me2Se suggests that the T* value of this compound is

Ea

-5 240\

v)

0

a

30 X

-1c

0 AROMATICS, r = 0 927 0 ALIPHATIC NON HBD }r=0975 A ALIPHATIC HBD

(8) The multiple linear regression equation with r* and a is A6 = 3.22 -

1 7 . W - 0.002a,with r = 0.974, which confirms that effects of type A hydrogen bonding by these aliphatic HBD solvents are negligible. (9) The d term is estimated through the equation d = 2(AXYZ)/[s(al) s(ar)], where AXYZ is the difference between values calculated through the aliphatic and aromatic regression equations at r* = 0.7 and s(a1) and s(ar) are the slopes of those regression equations. (IO) If the result for CHCll is included, r becomes 0.942. (11) We have commented earlier4 on the frequently observed seeming anomaly where weaker HBD solvents like CHCh form hydrogen bonds to weak HBA solutes; while stronger HBD solvents like MeOH or AcOH do not. The rationale is in the fact that the latter amphiprotic solvents are stronger HBA bases than the solute and achieve gieater stability by remaining self-associatedrather than by hydrogen bonding to the solute.

+

03

h

X

POLYCHLOROALIPHATIC

21x

057

\oy p'\ 290

-1-5

I 00

I

02

1 06

I 04

I 08

'

1 10

.* Figure 1. "Se NMR shifts of dimethyl selenide plotted against solvent =s values.

0.14-0.189 i.e.9 about like that for triethylamine (** = 0.14)9 and somewhat less dipolar than diethyl ether (T* = 0.27).

252 Inorganic Chemistry, Vol, 22, No. 2, 1983

Taft and Kamlet ~~

Q-53

I/-

r = 0 959 (ex 30)

r = 0.984

\

lex 30.3 1 )

76

2

-5

h D

Q

-10

1

031 A201 57% 2 I

I,.\ 21x

1

ll 1

0 ALIPHATIC NON HBD

A ALIPHATIC HBD 0 AROMATIC

,

0 ALIPHATIC NON HBD

,

A ALIPHATIC HBD

,

0 AROMATIC

X POLYCHLOROALIPHATIC

-15

8

x30

X

POLYCHLOROALIPHATIC

029

29

0 00

02

04

06

08

10 02

n*-0 42d

Figure 2. 77SeNMR shifts plotted against

T*

06

08

10

n*-026d

- 0.426.

Further, we have shown earlier2a*'2that the d term in eq 2 is related to the dipolarity/polarizability blend in the solvent effect on XYZ (expressed in terms of functions of the solvent refractive index and the bulk dielectric constant or molecular dipole moment), with d becoming more negative as the contribution of the polarizability term became smaller. The d value of -0.42 in eq 9 is among the largest negative values which we have yet encountered, which is consistent with the very weakly dipolar solute, Me2Se, causing relatively little polarization in cybotactic solvent molecules. These observations are in accord with the findings of Valeev et al., who, despite their failures at correlation, concluded that the predominating role is played by the polarity of the solvent. 12qe NMR Shifts of Dimethyl Telluride The same Russian workers have also reported solvent effects on the lz5Te N M R shifts of (CH3)2Te, again in 5 mol ?% solutions and against neat Me2Te as an external standard.I3 Their results in 17 solvents for which the solvatochromic parameters are known are also assembled in Table I. We have carried out the same types of correlations as above and, as before, the result for chloroform (solvent 30) was out of line." Also out of line in this case, for reasons that are not now evident to us, was the result for nitrobenzene (31). The other 15 solvents fit eq 1 and 2 rather well. As before, a plot against a* clearly showed the data points for the nonchlorinated aliphatic HBD and non-HBD solvents to be collinear, again indicating that type A hydrogen-bonding effects were negligible. The regression equation for the aliphatics is given by eq 10, with n = 7, r = 0.991, and sd = 1.6 ppm. For A6('25Te NMR) = 10.5 - 35.57P ppm

04

(10)

(12) Abboud, J-L,M.; Kamlet, M. J.; Taft, R. W., submitted for publication in J. Am. Chem. SOC.

(13) Kalabin, G. A.; Valeev, R. B.; Kushnarev, D. F. Zh. Org. Khim. 1981, 17, 947.

Figure 3. Iz5TeNMR shifts of dimethyl telluride plotted against - 0.266.

T*

the six aromatic solvents (excl31), the regression is given by eq 11, with r = 0.931 and sd = 2.2 ppm. From eq 10 and 11, A6('25Te NMR) = 15.4 - 30.2a* ppm (1 1) the d term was calculated to be -0.26,9 and the all-solvents correlation with T* - 0.266 is given by eq 12, with n = 15 (excl A6(IZ5TeNMR) = 8.4 - 32.6(a* - 0.266) ppm (12)

31 and 32),r = 0.984,14and sd = 1.9 ppm. A plot of As(expt1) against T * - 0.266 is shown in Figure 3. From eq 10 and 12 and the A6 value of 0.0 for neat Me2Te, the a* value for dimethyl telluride, acting as solvent, is calculated to be 0.26-0.30, Le., about as dipolar as diethyl ether. The greater dipolarity of Me2Te compared with that of Me2Se is consistant with the smaller -d value in eq 12 compared with that in eq 9. The more dipolar solute, Me2Te, induces the greater solvent polarizability contribution to the dipolarity/ polarizability blend. That NMR spectra of (CH3)2Seand (CH3),Te are not influenced by the HBD acidity of most protic solvents" indicates that they are very weak HBA bases. In a continuation of the progression from @ = 0.47 for diethyl ether,2s5and @ = 0.25-0.29 for diethyl sulfide (estimated from pKHB= 0.1 1 and the 4-fluorophenol I9FNMR value A = 1.10 ppmI5 and equations relating these properties to p5b),such weak HBA basicity properties for dialkyl selenides and tellurides are not unexpected. UV/Visible Spectrum of W(C0)Jphen) Solvatochromic behavior remarkably similar to that above (i.e., a large -d6 term and an insignificant dependence on the (14) If the result for nitrobenzene is included, the r value becomes 0.952; if

the CHCI, result is also included, r = 0.904. (15) Gurka, D.; Taft, R. W. J. Am. Chem. SOC.1969, 91, 4794.

Inorganic Chemistry, Vol, 22, No. 2, 1983 253

Linear Solvation Energy Relationships

} r = 0.978

0 ALIPHATIC

A ALIPHATIC HBD

0 ALIPHATIC

A ALIPHATIC HBD X

POLYCHLOROALIPHATIC

O/

I

0 AROMATIC

/"

X

POLYCHLOROALIPHATIC

V BENZYL ALCOHOL

105

7 oex 41

0 A104

r = 0978

04

08

06

02

10

04

Figure 4. vmx(I) plotted against solvent

T*

values.

Figure 5. umax(1) plotted against

-

a value of self-associated HBD solvents) is shown by a completely unrelated property, the metal ligand band in the electronic spectrum of tungsten tetracarbonyl 1,lophenanthroline (I), W(CO),(phen). Values of v-(I) reported by Burgess, Chambers, and Haines16 in the 18 solvents for which the solvatochromic parameters are known are included in Table I. All correlations discussed below will exclude an obviously out-of-line point for cyclohexanone (solvent 41). A plot of vmax(I)against T * is shown in Figure 4. It is seen in the figure that, as with the earlier examples, linear correlation is quite poor if all solvents are considered together (the r value is 0.706). If, however, the aliphatic and aromatic solvents are considered separately, good linear regressions are observed for the individual families. It is also seen that the data points for nitromethane (32), ethanol (104), and methanol (105) are collinear with those for the aliphatic non-HBD solvents and the data point for benzyl alcohol (109) is collinear with those for the aromatic non-HBD solvents, which is strong evidence that these HBD solvents do not hydrogen bond to I (or, far less likely, that they do hydrogen bond but the hydrogen bonding does not affect the spectrum). The correlation for the non-chlorinated HBD and non-HBD aliphatic solvents is given by eq 13, with n = 9, r = 0.978,''

vmx(I) = 103(18.21 + 3.821r*) cm-'

(1 3)

and sd = 0.17 X lo3 cm-'. The correlation for the aromatic solvents, including benzyl alcohol, is given by eq 14, with n

vmx(I) = 103(17.05

06

08

10

n*-0 30d

n*

+ 3.781r*) cm-'

?r*

- 0.306.

correlation according to eq 2 (excl CH2C12and CHC13,which probably hydrogen bond weakly to 1) is given by eq 15, with vmax(I)= 103[18.18

+ 3.85(r* - 0.306)] cm-'

(15)

n = 15, r = 0.978,'' and sd = 0.17

X lo3 cm-'. A plot of vmx(I) against r * - 0.306 is shown in Figure 5, where can be seen the weak bathochromic effect of hydrogen bonding by CHzClz(21) and CHC1, (30) (probably to the C 4 ligands)'* and the satisfactory linear regression for the other solvents. We have recently discussed the problems involved in determining r * values of HBD solvents (primarily because of difficulties in finding non-HBA indicators that allow us to unravel dipolarity/polarizability from hydrogen-bonding eff e c t ~ ) . For ~ ~ this reason eq 14 and 15 are of particular interest to us in that they are the first independent confirmation of the a * value of benzyl alcohol. Concluding Remarks We have now carried out sufficient correlations according to eq 2, 4, and 6 to arrive at some very rough preliminary generalizations regarding the sign and magnitude of the d coefficient. Thus, the d value of 0.0 for r 1r* transitions of uncharged indicators (all of which are shifted bathochromically with increasing solvent dipolarity) appears very nearly to represent a limiting condition. We have encountered only one unambiguous example of a positive d term, the fluorescence spectrum of the merocyanine 11, for which d = +0.08.20

-

(14)

= 6,r = 0.966, and sd = 0.21 X lo3 cm-'. From eq 13 and 14, the d term is calculated to be -0.30,9 and the all-solvents

(18) We are very hesitant in suggesting this rationale because of the uncertainty of the 6 value of 0.5 for the polychloro aliphatic solvents. The results could also be rationalized by a larger 6. Further, hydrogen

(16) Burgess, J.; Chambers, J. C.; Haines, R. I. Transition Met. Chem. (Weinheim, Ger.) 1981, 6, 145. (17) If the result for cyclohexanone is included, r becomes 0.960 for eq 13 and 0.955 for eq 15. If CH2C12and CHCIJ are included, r becomes 0.952 for eq IS.

bonding to carbonyl should withdraw electrons and hence destabilize ligand transition. This should lead to the excited state in a metal a hypsochromic effect of hydrogen bonding rather than the observed bathochromic effect. (19) Chawla, 8.; Pollack, S. K.; Lebrilla, C. B.; Kamlet, M. J.; Taft, R. W. J. Am. Chem. SOC.1981, 103, 6924. (20) Kamlet, M.J.; Taft, R. W. Pol. J . Chem., in press.

-

Inorg. Chem. 1983, 22, 254-266

254

Electronic ?r ?r* transitions of charged indicators, which are shifted bathochromically (e.g., Cnitrophenoxide ion),21and all electronic spectra, which are shifted hypsochromically (i.e., all spectra wherein the magnitude of the dipole decreases in the electronic excitation) have relatively large negative d terms; we have encountered values ranging from -0.15 to -0.40. The same applies to 'H, 13C, lSN, 19F,29Si,and l19Sn NMR shifts and coupling constants (as well as the 77Seand 125Teshifts reported here). As has been mentioned, values of d have generally been lower for infrared r, results, generally ranging from -0.05 to -0.15.

In sharp contrast with the trend reported here for the NMR results, -d values for free energies of transfer between solvents seem to decrease with decreasing solute dipolarity. We have reported22values near nil for toluene and dioxane solutes, -0.09 for the 2-butanone solute, -0.20 for the nitromethane solute, and -0.17 for the Et4N+I-ion pair. As concerns reaction rates, we have encountered d values of -0.09 to -0.17 for various types of Menschutkin reactions, but we have been unable to relate d to transition-state dipolarity or structure. Thus, we have encountered both larger and smaller -d values for reactions with less dipolar transition states. Acknowledgment. The work by R.W.T. was done in part under a grant from the Public Health Service. The work by M.J.K. was done under Naval Surface Weapons Center Foundational Research Task No. IR-210. Registry No. I, 14729-20-5;77Se,14681-72-2;Me2Se, 593-79-3; I2'Te, 14390-73-9; Mn2Te, 593-80-6.

(21) Abboud, J. L. M.; Taft, R. W.; Kamlet, M.J. Bull. Chem. Soc. Jpn. 1982, 55, 603.

(22) Kamlet, M. J.; Carr, P. W.; Taft, R. W.; Abraham, M. H. J. Chem. Soc., Faraday Trans. I 1982, 78, 1689.

IIa

IIb

-

Contribution from Bell Laboratories, Murray Hill, New Jersey 07974

Electronic Structure and Bonding in the Disulfur and Diselenium Complexes [M(X,)(PH,),]+ (M = Rh, Ir; X = S, Se) A. P. GINSBERG,* J. H. OSBORNE,' and C. R. SPRINKLE

Received April 28, I982

SCF-Xa-SW calculations have been carried out on the model compounds [M(X2)(PH3)4]+(M = Rh, Ir; X = S, Se) in order to investigate the electronic structure and bonding in complexes of side-on-bonded disulfur and diselenium. In agreement with experiment, the calculations predict an X-X bond order of about 1 and reveal that the M-X2 covalent interaction increases along the sequence RhSe2 < I r h < RhS2< IrS2. Relativisticeffects partly account for the increase in interaction from Rh to Ir. The side-on-bonded S2 and Se2 groups are best described as molecules with excited configuration ...(pU ) ~ ( A ) ~ ( T *The ) ~ .M-S2 or M-Se2 bond consists of in-plane A overlap of an S2 or Se2T,,* orbital with a metal px + d, hybrid of predominantly px character and of a overlap of a metal dz2 + pz hybrid with S2or Se2 rIIand pa orbitals. Optical spectra of [M(X,)(dppe),J+ in EPA glasses at liquid-nitrogen temperature are reported for the 300-800-nm region. All bands are assigned, and acceptable agreement is found between observed and calculated transition energies. The HOMO LUMO transition occurs as a weak band in the 530-700-nm region; it is primarily intraligand (S2or Se2) in character. The lowest energy strong band is associated with the transition from the M-X2 r-bonding orbital to the LUMO; it shifts to higher energy from M = Rh to M = Ir on account of the increase in M-X2 interaction.

-

Introduction Disulfur and diselenium can bond to metals in a variety of bridging geometries and also as terminal groups in the side-on manner2 My: X'

Side-on-bonded S2and Se, complexes are, of course, analogues of the well-known side-on-bonded dioxygen complexes, but unlike the case of the dioxygen complexes, their chemistry has been relatively little studied, and there is no reliable information on their electronic structure. Recent work, especially the preparation of novel new disulfur complexes3 and the

(5Department 1981 summer research student at Bell Laboratories. Present address: of Chemistry, University of Washington, Seattle, WA.

discovery of the addition of low-valent-metalcomplexes across the S-S and Se-Se bond,44 indicates that further studies on the chemistry of side-on-bonded S2and Se, are likely to yield interesting results. As a guide for future work, and for interpretation of experimental results, it would be valuable to have detailed information about the electronic structure and bonding in the

group. This paper provides such information from an SCFXa-SW investigation of the electronic structure of the complexes [M(X,)(PH,),]+ (M = Rh, Ir; X = S , Se). The complexes [M(X2)(PH3)4]+are models for the compounds [M(X,)(L-L),]+ (M = Rh, Ir; X = S , Se; L-L = dppe, dmpe), and the calculations are in good agrement with

(2) For a summary of types of metal-disulfur geometries see A. Muller and M. Jagermann, Inorg. Chem., 18,2631 (1979). (4) (3) Exampla: (NH,)*[(S*)*MO(S~)*MO(S~~] by A. MIUler, W. Nolte, and B. Krebs Angew. Chem., Int. Ed. Engl., 17, 279 (1978); (NH& [Mo3S(S2),] by A. Miiller, S . Sarkar, R. G.Bhattacharyya, S . Pohl, (5) and M. Dartmann, ibid., 17, 535 (1978); [(CH3)4N]2[Mo202S2(S2)2] by W. Rittner, A. Muller, A. Neumann, W. Bather, and R. C. Sharma, (6) ibid., 18, 530 (1979).

0020-166918311322-0254%01,5010

D. Seyferth, R. S . Henderson, and M. K. Gallagher, J . Organomet. Chem. 193, C75 (1980). D. A. Lesch and T. B. Rauchfuss, J. Organomet. Chem., 199, C6 (1980). A. P. Ginsberg,W. E. Lindsell, C. R. Sprinkle, K. W. West, and R. L. Cohen, Inorg. Chem., 21, 3666 (1982).

0 1983 American Chemical Society